Works & Aesthetics

Blog Posts

Signal Detection Theory


Signal Detection Theory

What is signal Detection Theory?

Signal detection theory (SDT) is a framework for understanding accuracy that makes the role of decision processes explicit. To do so, the theory also takes a stand on the way in which the relevant information is represented by the observer, identifying some aspects of the representation with sensitivity, or inherent accuracy, and others with response factors. 


Brief History

Signal Detection theory was developed in the 1950’s for mathematical statistics and electronic communication purposes. More specifically,  it was developed to monitor the performance of radars, which must detect signals against other stimuli/noise that can either exist in the background or be as strong as the signal. 


Signal detection theory is very simple to explain. Let’s use as an example the usage of signal detection theory in radars. 

When radars correctly detect a signal (Weak or strong) that is called a “hit”. If a radar identifies a stimuli as signal where in fact is noise, then this is a case of “false alarm”. Respectfully, when a radar identifies a stimuli as noise, when in fact, it is an actual signal, then that is case of a “miss”. Finally, when radars identify a stimuli as noise, and it is actual noise, then that is a case of correct rejection. 


Signal detection theory is not only used in mathematics and engineering. Nowadays it is widely used in the social sciences, such as psychology, and in everyday life. 

In diagnosing cancer with imperfect methods, is it better to fail to detect a tumor or to detect one that is not present? Which way should an eyewitness lean—toward failing to report recognizing someone who has perpetrated a crime, or toward accusing someone who was not the criminal?

That such choices are possible makes clear the importance of decision processes in perception. Signal detection theory (SDT) is a framework for understanding accuracy that makes the role of decision processes explicit.

SDT can be applied to any binary decision making situation where the response of the decision maker can be compared to actual presence or absence of the target.